> Yes, I know that. But in the context of the science I'm doing, a
> pure POSIX-only system is not useful.
I think several people have made the comment over the last little
while in relation to both unix and POSIX like systems, the there are
enough undefined gaps in POSIX that just because something is
POSIX-compliant, it doesn't make it useful.
I think the NT POSIX subsystem would probably fall into this
category, it may be (or have been) POSIX compliant, but its
sufficiently hard to actually do anything useful with it...
> A Unix system, however, is useful. And one of my points is that
> most (all?) existing Unix systems have EFAULT, and we should avoid
> diverging from that behaviour.
I agree completely. Linux != POSIX, Solaris != POSIX, and thats
perfectly acceptable.
> I have no problem with optional SEGV behaviour. But to abolish
> EFAULT is a Bad Thing.
Again, I agree.
> Solaris and every other Unix I've come across.
One thing I wonder, how does Solaris (2.5.1 and earlier) handle bogus
pointers to things like getpeername, etc - which are really libc
wrappers?
> But since most (all?) Unix systems have EFAULT, is it wise to
> deviate from that? Espectially considering that EFAULT is actually
> *useful* and is *used in real life*.
I think EFAULT is useful and acceptable behavior. Sure, it breaks the
syscall/function call transparency arguments - but I don't see how
having a higher degree of transparency is useful in practice (yes, in
_theory_ it might feel better, but then you may as well write all
your OS in scheme if thats all your interested in).
-cw
(I'm thinking this thread should probably be killed soon. These
bloody devfs people are shit stiring again...)
-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.rutgers.edu
Please read the FAQ at http://www.altern.org/andrebalsa/doc/lkml-faq.html